Robotics Simulation and Evaluation PhD Intern - Summer 2027
Core
Develop novel approaches to agentic robot policies, world action models, robot policy training and evaluation, and sim-to-real transfer using GPU-accelerated simulation.
Role type
PhD intern in robotics simulation and evaluation
Builds
NVIDIA's robotics and simulation platforms
Domain
Robotics, Machine Learning, GPU-accelerated simulation
Deliverable
production ML models
Required skills
Python, deep learning frameworks (PyTorch, JAX), robot learning, sim-to-real transfer, research code development, experimental design
Preferred skills
reinforcement learning, imitation learning, large vision-language-action models, GPU-accelerated physics simulation, real robot hardware deployment, large-scale multi-GPU training
Technologies
PyTorch, JAX, PhysX, Isaac Gym, Isaac Lab, Newton, MJX
Responsibilities
Scope scientifically novel projects relevant to the robotics roadmap; Design, implement, and evaluate new robot learning methods; Run rigorous experiments including baselines and ablation studies; Write clean, well-documented research code; Collaborate to transfer research into products; Deliver results as prototypes, patents, or publications
Seniority
PhD Intern
